Received: 16 July 2017 | Revised: 21 August 2017 | Accepted: 4 September 2017 DOI: 10.1111/ajt.14503 ORIGINAL ARTICLE Defining a microRNA- mRNA interaction map for calcineurin inhibitor induced nephrotoxicity Christopher J. Benway1,2 | John Iacomini1,2,3,4 1Department of Integrative Physiology and Pathobiology, Tufts University School of Calcineurin inhibitors induce nephrotoxicity through poorly understood mechanisms Medicine, Boston, MA, USA thereby limiting their use in transplantation and other diseases. Here we define a mi- 2 Graduate Program in Genetics, Sackler croRNA (miRNA)- messenger RNA (mRNA) interaction map that facilitates exploration School of Graduate Biomedical Sciences, Tufts University School of Medicine, Boston, MA, into the role of miRNAs in cyclosporine- induced nephrotoxicity (CIN) and the gene USA pathways they regulate. Using photoactivatable ribonucleoside- enhanced crosslinking 3 Graduate Program in Immunology, Sackler and immunoprecipitation (PAR- CLIP), we isolated RNAs associated with Argonaute 2 School of Graduate Biomedical Sciences, Tufts University School of Medicine, Boston, MA, in the RNA-induced silencing complex (RISC) of cyclosporine A (CsA) treated and con- USA trol human proximal tubule cells and identified mRNAs undergoing active targeting by 4Tufts University School of Medicine, Boston, MA, USA miRNAs. CsA causes specific changes in miRNAs and mRNAs associated with RISC, thereby altering post-transcriptional regulation of gene expression. Pathway enrich- Correspondence John Iacomini ment analysis identified canonical pathways regulated by miRNAs specifically follow- Email: [email protected] ing CsA treatment. RNA- seq performed on total RNA indicated that only a fraction of total miRNAs and mRNAs are actively targeted in the RISC, indicating that PAR- CLIP more accurately defines meaningful targeting interactions. Our data also revealed a role for miRNAs in calcineurin- independent regulation of JNK and p38 MAPKs caused by targeting of MAP3K1. Together, our data provide a novel resource and unique in- sights into molecular pathways regulated by miRNAs in CIN. The gene pathways and miRNAs defined may represent novel targets to reduce calcineurin induced nephrotoxicity. KEYWORDS basic (laboratory) research/science, genomics, immunosuppression/immune modulation, kidney failure/injury, molecular biology: micro RNA, molecular biology: mRNA/mRNA expression, translational research/science 1 | INTRODUCTION causing glomerular and vascular disease, abnormal tubular function, and increased blood pressure.6 Acute and chronic CIN can only be Calcineurin inhibitors are a cornerstone of immunosuppression for diagnosed by invasive renal biopsy, usually after significant kidney transplant patients and have dramatically reduced rejection rates be- damage has occurred. There a significant gap in our knowledge in the cause of their ability to prevent T cell activation.1 This ability has also diagnosis, prevention and treatment of CIN driven by an incomplete led to their use in the treatment of autoimmune diseases.2,3 However, understanding of its molecular basis. While CIN appears to result from short- or long- term use of calcineurin inhibitors such as CsA and processes that cause apoptosis of renal tubular cells, generation of re- tacrolimus is associated with nephrotoxicity.1,4 Calcineurin inhibitor active oxygen species, and epithelial- mesenchymal transition (EMT),7- induced nephrotoxicity (CIN) can be acute or chronic.5 While acute 9 how these processes are regulated is not completely understood. CIN is reversible following drug withdrawal, chronic CIN is irreversible Developing an understanding of the gene pathways involved in CIN 796 | © 2017 The American Society of Transplantation amjtransplant.com Am J Transplant. 2018;18:796–809. and the American Society of Transplant Surgeons BENWAY AND IACOMINI | 797 and how they are regulated could offer novel prevention and treat- 2.4 | AGO2- PAR- CLIP ment strategies to reduce toxicity allowing better usage of these po- tent immunosuppressive drugs. PAR- CLIP was performed as previously described.49,50 MicroRNAs (miRNAs) are a class of small, noncoding RNAs that post- transcriptionally regulate the expression of more than 60% 2.5 | PAR- CLIP bioinformatic analysis and ingenuity of protein coding genes in humans10-16 and control cellular func- pathway analysis tion in several disease processes.17-29 We previously demonstrated that miRNAs play an important role in controlling acute renal injury Details are provided in Supplemental Materials and Methods. resulting from ischemia and reperfusion.30 It is now apparent that the miRNAs play a key role in regulating many aspects of renal in- 2.6 | Western blot assay jury.31-45 Because miRNAs regulate gene expression, understand- ing how their expression is altered throughout the progression of Cell cultures were lysed in RIPA buffer supplemented with phos- a disease lends itself to the discovery of new pathways involved in phatase and protease inhibitors. SDS- PAGE was performed using disease progression, development of phenotyping assays and bio- 4- 12% Novex NuPage Bis- Tris gels in MOPS buffer under reducing markers, and identification of novel therapeutic targets. We hy- conditions (Thermo Fisher). Monoclonal antibodies used for were pothesized that identification of genes undergoing active targeting anti- MAP3K1/MEKK1 (Clone 2F6, Novus Biologicals, Littleton, CO) by miRNAs would allow us to identify pathways altered due to CIN and anti- GAPDH (Clone 71.1, Sigma- Aldrich). and define global changes in gene regulation that contribute to dis- ease and potential novel targets for therapeutic intervention. To 2.7 | Real- time quantitative PCR test our hypothesis, we used crosslinking and immunoprecipitation (CLIP) methodology to identify miRNAs and mRNAs associated with PCR was performed as previously described46 after preparing cDNA Argonaute 2 (AGO2) protein in the RNA- induced silencing complex using the High Capacity cDNA Reverse Transcription Kit (Applied of CsA treated and control human proximal tubule cells and then Biosystems, Foster City, CA). TATA- binding protein (human TBP or used that information to define a miRNA- mRNA targetome for CIN. mouse Tbp) mRNA was used as normalization control. Relative quanti- This analysis revealed several pathways under regulation by miRNAs fication was calculated using the ΔΔCT method and assays were per- in CsA treated cells and an unexpected role for miRNAs in regulating formed in triplicate or quadruplicate.51 expression of MAPKs. 2.8 | Availability of data and material 2 | MATERIALS AND METHODS All raw sequencing data and associated processed PAR- CLIP, RNA- seq, and miRNA- seq datasets discussed in this publication have been 2.1 | Cell culture deposited in NCBI’s Gene Expression Omnibus and are accessible HK- 2 cells (ATCC, Manassas, VA) were cultured in serum- free ke- through GEO Series accession number GSE98670. ratinocyte medium (Thermo Fisher, Waltham, MA) supplemented with Epidermal Growth Factor 1- 53 and Bovine Pituitary Extract and 3 | RESULTS treated with 5 μg/mL CsA (Sigma- Aldrich, St. Louis, MO) or ethanol as a vehicle control as previously described.46 3.1 | PAR- CLIP defines miRNAs associated with the RISC complex in human proximal tubule cells treated 2.2 | Animal model of CsA nephrotoxicity with CsA Ten- to 12- week- old male C57BL/6J mice were purchased from The We performed AGO2- PAR- CLIP on human kidney proximal tubule Jackson Laboratory (Bar Harbor, ME) and housed under microisola- cells (HK- 2 cells) treated with CsA or vehicle as a control. HK- 2 cells tor conditions. All experiments were performed with approval of our were cultured in medium supplemented with 100 μM 4- thiouracil (4- Institutional IACUC. CsA nephrotoxicity was induced as previously SU) and then treated with CsA. The cells were then exposed to UV described.46-48 light (365 nm) to crosslink RNAs to associated RNA binding proteins. AGO2- associated RNAs (miRNAs and mRNAs) were then isolated by immunoprecipitation using AGO2- specific antibodies (Figure S1). 2.3 | RNA- Seq AGO2- associated RNAs were then purified, reverse transcribed, HK- 2 cells were treated for 48 hours with CsA (5 μg/mL) or vehicle amplified, and made into cDNA libraries and sequenced. PAR- CLIP control (0.1% ethanol). Four replicates were used for each condition. was performed in duplicate for each experimental condition and the Polyadenylated mRNAs were isolated and used to generate multi- resulting Illumina sequencing reads were pooled for maximal read plexed cDNA libraries for Illumina sequencing. depth. Raw sequencing clusters, trimmed reads and mapped reads are 798 | BENWAY AND IACOMINI A B C D E F FIGURE 1 Analysis of RISC- bound RNAs and miRNAs and mRNAs obtained by AGO2- PAR- CLIP in CsA- treated HK- 2 cells. (A) Mutational frequencies of all possible transitions in AGO2- PAR- CLIP sequences derived from vehicle (Veh) and CsA- treated HK- 2 cells. (B) Proportions of the top 20 miRNAs that were detected in AGO- PAR- CLIP libraries by alignment to miRNAbase v21 reference database. Below, log2 fold change of top 20 miRNAs detected in RISC complex after treatment with CsA. (C) Genomic distribution of crosslinking sites identified by AGO2- PAR- CLIP in vehicle and CsA- treated HK- 2 cells. (D) Log2 enrichment of AGO2 crosslinking sites in genomic regions based on normalization to relative region sizes in the reference genome. 3UTR: 3’ untranslated region, 5UTR: 5’ untranslated region, ncRNA: